Training data generation device, recording method, and inference device

Abstract

A training data generation device includes a computer, and a computer-readable storage medium. The computer is configured to: receive an input of an annotation for second image data obtained by imaging an observation target; reflect a result of the annotation in first image data that is related to the same observation target as the observation target of the second image data, the first image data having a different at least one of imaging mode and display mode from the second image data; and generate training data for creating an inference model by using the first image data and the result of the annotation reflected in the first image data, the first image data including image data of a plurality of images, and the second image data being image data of an image obtained by combining the plurality of images included in the first image data.

What is claimed is: 
     
         1 . A training data generation device comprising
 a computer, and   a non-transitory computer-readable storage medium,   the computer being configured to:   receive an input of an annotation for second image data obtained by imaging an observation target;   reflect a result of the annotation in first image data that is related to the same observation target as the observation target of the second image data, the first image data having a different at least one of imaging mode and display mode from the second image data; and   generate training data for creating an inference model by using the first image data and the result of the annotation reflected in the first image data,   the first image data including image data of a plurality of images, and   the second image data being image data of an image obtained by combining the plurality of images included in the first image data,   wherein the plurality of images are images captured by different imaging units at different positions,   wherein the image file includes out-of-screen coordinate information as the metadata, the out-of-screen coordinate information serving as the annotation coordinate information for the individual image data.   
     
     
         2 . The training data generation device according to  claim 1 , wherein the second image data is image data to be displayed on a display when the annotation is performed on the second image data. 
     
     
         3 . The training data generation device according to  claim 2 , wherein the different imaging mode is processing for improving visibility when displaying on the display. 
     
     
         4 . The training data generation device according to  claim 1 ,
 wherein the first image data is image data of a raw image, and   the second image data is image data of an image obtained by performing image processing on the raw image.   
     
     
         5 . The training data generation device according to  claim 4 , wherein the image processing includes at least one of gamma correction, white balance adjustment, color correction, noise reduction, and image enhancement processing. 
     
     
         6 . The training data generation device according to  claim 1 ,
 wherein the first image data is image data of an image captured in a state where the observation target is irradiated with special observation light, and   the second image data is image data of an image captured in a state where the observation target is irradiated with illumination light.   
     
     
         7 . The training data generation device according to  claim 1 , the computer is further configured to calculate coordinates of an image included in the first image data or the second image data. 
     
     
         8 . A recording method comprising:
 performing image combining processing to display, on a display, a plurality of pieces of image data obtained by imaging an observation target;   acquiring annotation coordinate information on a display image of composite image data displayed on the display;   adding the annotation coordinate information to individual image data constituting the composite image data displayed on the display;   creating an image file including the annotation coordinate information for the individual image data as metadata; and   recording the created image file,   wherein the plurality of images are images captured by different imaging units at different positions,   wherein the image file includes out-of-screen coordinate information as the metadata, the out-of-screen coordinate information serving as the annotation coordinate information for the individual image data.   
     
     
         9 . An inference device comprising:
 a non-transitory computer-readable storage medium storing a trained model that has been trained using training data; and   a computer configured to:   receive an input of an annotation for second image data obtained by imaging an observation target;   reflect a result of the annotation in first image data that is related to the same observation target as the observation target of the second image data, the first image data having a different at least one of imaging mode and display mode from the second image data;   generate training data for creating an inference model by using the first image data and the result of the annotation reflected in the first image data; and   perform inference processing on first image data based on the trained model that has been trained using the training data generated,   the first image data including image data of a plurality of images, and   the second image data being image data of an image obtained by combining the plurality of images included in the first image data,   wherein the plurality of images are images captured by different imaging units at different positions,   wherein the image file includes out-of-screen coordinate information as the metadata, the out-of-screen coordinate information serving as the annotation coordinate information for the individual image data.